Commentary

The first tab displays a table with the characteristics of patients with diagnostic codes for viral bronchiolitis.

Patients in the “Primary Bronchiolitis Diagnoses” cohort have a primary admission diagnosis of viral bronchiolitis. Patients in the “All Bronchiolitis Diagnoses” cohort have an ICD9 or ICD10 code for viral bronchiolitis associated with admission as a primary or secondary diagnosis code. The cohort in the last column is identified using the All Patient Refined Diagnosis Group, version 32 (APR-DRG v32) for viral bronchiolitis using 3M software (3M, St. Paul, MN).

The second tab shows show the percentage of admissions within a given discharge year that had CCC.

The third tab shows the relative composition of CCCs over time. CCCs are shown as a proportion of the total number of CCCs because a given admission may have more than one CCC (e.g a patient may be admitted with both a respiratory and a renal CCC).

Commentary

The first four tabs display the number of annual admissions, unique patients, ICU admissions, and admissions requiring ECMO over time. These graphs display the number of encounters for the selected PHIS hospitals among all diagnoses (dark gray) compared to bronchiolitis (light gray). The bars represent the absolute number of admissions, and the line (right axis) shows the trend in the proportion of bronchiolitis diagnoses over time.

These graphs show that the proportion of patients with a primary admission diagnosis of bronchiolitis is relatively constant over time, but that the number of patients with any admission diagnosis of bronchiolitis is increasing over the past decade.

Bronchiolitis represents a gradually increasing proportion of ICU admissions over time across all cohorts. The proportion of patients with any bronchiolitis diagnosis receiving ECMO is increasing over time, though the proportion receiving ECMO is stable in the primary diagnosis and APR-DRG cohorts.

The last tab shows the proportion of admissions with bronchiolitis admitted to the ICU over time. The first three graphs show the bronchiolitis cohorts. The last graph shows all admissions to the PHIS hospitals among children under two years old (with any diagnosis, not just bronchiolitis) for reference.

Commentary

The first tab shows the proportion of patients receiving invasive (dark gray) and non-invasive (light gray) ventilation over time. The proportion of patients receiving non-invasive ventilation has significantly increased across all cohorts. The proportion of patients receiving invasive mechanical ventilation has not changed significantly in any cohort.

The second tab shows the use of invasive (dark gray) and non-invasive (light gray) ventilation by hospital. There was no significant correlation in any cohort.

The third tab shows the number of patients with bronchiolitis admitted to the ICU receiving no ventilatory support (lightest gray), non-invasive ventilation (medium gray), and invasive mechanical ventilation (darkest gray).

The fourth tab shows the percentage of patients with bronchiolitis admitted to the ICU receiving no ventilatory support (lightest gray), non-invasive ventilation (medium gray), and invasive mechanical ventilation (darkest gray).

Commentary

The first tab shows the hospital length of stay over time for patients with bronchiolitis and the third tab shows the ICU length of stay.

The second tab and the fourth tab show the hospital length of stay and the ICU length of stay, respectively, for patients with bronchiolitis grouped by their maximum level of ventilator support: none (light gray), non-invasive (medium gray), or invasive (dark gray).

Note that the plots of ICU length of stay include only patients admitted to the ICU. This cohort has a longer hospital length of stay compared to the cohort not admitted to an ICU, which is why the median ICU length of stay for this cohort is approximately the same as the median hospital length of stay for the cohorts of all patients with bronchiolitis.

Overall hospital length of stay and ICU length of stay are decreasing for patients with oxygen support only and patients with non-invasive ventilatory support. Length of stay has not significantly changed over time for patients who receive invasive ventilation.

Commentary

These graphs show cost of care over time. The first tab and the second tab show the annual cost adjusted for the Bureau of Economic Analysis Gross Domestic Product (GDP) and CMS National Healthcare Expenditures per Capita (NHEC), respectively.

The third tab and the fourth tab show the per-admission cost adjusted for GDP and NHEC, respectively.

The fifth tab and the sixth tab show the per-admission cost adjusted for GDP and NHEC, respectively, grouped by the level of ventilatory support.

For all grouped graphs, costs are grouped by level of ventilatory support: no ventilatory support (lightest gray), non-invasive ventilatory support (light gray), invasive ventilatory support (dark gray), and both invasive and non-invasive ventilatory support (darkest gray).

All graphs adjust for Centers for Medicare & Medicaid Services (CMS) cost-of living index adjust for cost-of-living in each hospital’s zip code.

The graphs adjusted for the Bureau of Economic Analysis Gross Domestic Product (GDP) account for non-specific inflation and are expressed in 2010 dollars. The graphs adjusted for CMS National Healthcare Expenditures per Capita costs account for healthcare-specific inflation and are expressed in 2010 dollars.

Adjusted costs have significantly decreased over time for patients receiving non-invasive ventilation, invasive ventilation or both during hospitalization.

Commentary

Results of multivariable logistic regression models with an outcome of invasive mechanical ventilation are presented. Models were developed by including race/ethnicity, age, sex, insurance status, complex chronic condition flags and discharge year. Automated backward stepwise selection was performed to minimize Akaike Information Criteria. Discharge year was forced into the model as it was a variable of interest. The mixed effects models contain random intercepts for hospital region.

The first tab shows the results for patients with a primary encounter diagnosis of bronchiolitis.

The second tab shows results for patients with any encounter diagnosis of bronchiolitis (primary or secondary).

The third tab shows the results for patients within the 3M APR-DRG v32 group for bronchiolitis.

Commentary

Results of multivariable logistic regression models with an outcome of non-invasive ventilation are presented. Models were developed by including race/ethnicity, age, sex, insurance status, complex chronic condition flags and discharge year. Automated backward stepwise selection was performed to minimize Akaike Information Criteria. Discharge year was forced into the model as it was a variable of interest. The mixed effects models contain random intercepts for hospital region.

The first tab shows the results for patients with a primary encounter diagnosis of bronchiolitis.

The second tab shows results for patients with any encounter diagnosis of bronchiolitis (primary or secondary).

The third tab shows the results for patients within the 3M APR-DRG v32 group for bronchiolitis.

Commentary

Results of multivariable logistic regression models with an outcome of icu admission are presented. Models were developed by including race/ethnicity, age, sex, insurance status, complex chronic condition flags and discharge year. Automated backward stepwise selection was performed to minimize Akaike Information Criteria. Discharge year was forced into the model as it was a variable of interest. The mixed effects models contain random intercepts for hospital region.

The first tab shows the results for patients with a primary encounter diagnosis of bronchiolitis.

The second tab shows results for patients with any encounter diagnosis of bronchiolitis (primary or secondary).

The third tab shows the results for patients within the 3M APR-DRG v32 group for bronchiolitis.

Commentary

The first tab shows the proportion of admissions with bronchiolitis admitted to a single-center ICU over time. The first graph shows patients with a primary encounter diagnosis of bronchiolitis. The second graph shows all admissions to the single-center among children under two years old (with any diagnosis, not just bronchiolitis) for reference.

The second tab shows the number and percentage of patients with bronchiolitis admitted to the ICU receiving no ventilatory support (lightest gray), non-invasive ventilation (medium gray), and invasive mechanical ventilation (darkest gray).

The third tab shows the modified electronic PELOD-2 score among patients admitted to the single-center ICU with a primary encounter diagnosis of bronchiolitis. Patients receiving no ventilation support (high flow nasal cannula) and non-invasive ventilation were grouped together (light gray) and compared with patients receiving invasive ventilation (dark gray).

The fourth tab shows the modified electronic PRISM-IV score among patients admitted to the single-center ICU with a primary encounter diagnosis of bronchiolitis. Patients receiving no ventilation support (high flow nasal cannula) and non-invasive ventilation were grouped together (light gray) and compared with patients receiving invasive ventilation (dark gray).

Severity of illness data were available from 2010 to 2018.